Voice AI Agents
Real-time voice assistants built on LiveKit — for support, booking, and lead qualification. This site runs one; try the floating widget.
Learn more →Making AI feel natural, not futuristic.
I'm Ellis Velandia, a full stack developer in Bogotá. I build production AI agents — voice agents that answer calls, chatbots grounded in your data, and automation that quietly runs your workflows. Without the cold, techy feel.
English & Spanish · reply within 48h · no sales pitch
Ellis — the human behind the agents
What I do
Four ways I put AI to work
Each one ships as a production system with guardrails — not a demo that impresses once and breaks.
Chatbots & Copilots
Chatbots grounded in your own data with RAG, tool calling, and guardrails — powered by Claude, GPT, or Gemini.
Learn more →Workflow Automation
Agents that draft content, triage email, sync CRMs, and run scheduled pipelines — like the daily blog automation on this site.
Learn more →AI Integration
Semantic search, summarization, extraction, and MCP tool servers added to your existing React/Next.js product.
Learn more →Proof
The voice widget you can talk to on this site, and the blog that writes itself daily in two languages — that's the same agent architecture I ship to clients.
Ellis Velandia · see it live on the blog and in projectsHow we'll work
A collaborative process
- 1
Discover
A free 30-minute call. We find the one job your agent must do well — and whether an agent is even the right tool.
- 2
Design
Architecture, tools, data grounding, and guardrails. You get a written proposal with fixed pricing.
- 3
Build
Short iterations you can try as they land. An evaluation set from week one — not vibes.
- 4
Deploy
Edge deployment, monitoring, a human handoff path, and support after launch.
FAQ
Common questions
What is an AI agent?
An AI agent is software that uses a large language model (LLM) to reason, use tools, and complete tasks autonomously — answering customers by voice or chat, processing documents, or automating multi-step workflows without human intervention.
How much does it cost to build an AI agent?
A focused MVP agent (one channel, one workflow) typically starts in the low four figures USD and ships in 2–4 weeks. Complex voice agents or multi-tool automations scale with scope. Every project starts with a free scoping call.
Which technologies do you use to build AI agents?
Claude and GPT model APIs, LiveKit for real-time voice, Next.js and TypeScript for the app layer, Supabase (Postgres + pgvector) for data and retrieval, and Cloudflare Workers for low-latency edge deployment.
Can an AI agent work in Spanish and English?
Yes. As a native Spanish speaker working in English daily, Ellis builds fully bilingual agents — voice and chat — that switch language automatically based on the user.
How do I hire Ellis Velandia to build an AI agent?
Book a free intro call through the contact page. You will get a scoped proposal with timeline and fixed pricing within a few days.
Have a workflow that eats your team's hours?
Tell me about it on a free intro call. You'll get a scoped proposal with a timeline and fixed pricing — usually within a few days.